ElevenLabs

Stream meeting audio to your ElevenLabs agent and play its responses back into the meeting.

Configure your agent's prompt, voice, and tools in ElevenLabs. Recall.ai opens an authenticated conversation over WebSocket, sends the meeting audio mix, and plays the agent's replies into the meeting.

Setup

1. Prepare your ElevenLabs agent

Create or open an agent in ElevenLabs, configure its conversation, and copy its agent ID. Create an API key with permission to access that agent.

Configure the agent to use PCM audio for both input and output, such as pcm_16000. Recall.ai supports pcm_8000, pcm_16000, pcm_22050, pcm_24000, pcm_44100, and pcm_48000, and resamples as needed. Other formats, including μ-law, are not supported by this integration.

2. Add your credentials to Recall.ai

Open Agents → ElevenLabs in the Recall region where you create bots:

Select the workspace used by your Recall API key, then save the ElevenLabs key and select the Data residency region that matches your ElevenLabs agent: United States, European Union, or India.

The saved key and ElevenLabs region are shared with ElevenLabs transcription. This region is configured separately from the Recall region where you create bots. Changes apply to new connections.

Quickstart

Call Create Bot with your meeting URL and agent ID. Replace agent_YOUR_AGENT_ID with the ID from ElevenLabs.

We recommend setting the relevant meeting platform in variant to web_4_core when pairing this audio agent with a video provider or a kind: "webpage" source. See bot sizing guidance.

{
  "meeting_url": "MEETING_URL",
  "output_media": {
    "audio": {
      "kind": "elevenlabs",
      "config": {
        "agent_id": "agent_YOUR_AGENT_ID"
      }
    }
  }
}

Once the bot joins the meeting, Recall.ai obtains a signed conversation URL using the saved credentials, connects to the agent, and exchanges audio.

For recording the agent's voice, see Recording meeting agents. For response controls, see the Meeting agents FAQ.

Additional configurations

The output_media.audio.config object supports:

FieldRequiredDescription
agent_idYesYour ElevenLabs agent ID: 1–256 ASCII letters, digits, underscores, or hyphens.
dynamic_variablesNoA JSON object containing values for dynamic variables used by the agent. Recall.ai adds the reserved bot variables described below.
conversation_config_overrideNoA JSON object matching conversation_config_override in ElevenLabs' conversation_initiation_client_data schema. Enable the relevant overrides in your agent's security settings. Audio mode must remain enabled.

The API key and data residency region are read from saved credentials. api_key, credential_id, and data_residency are not accepted in this config.

Personalize a conversation

Use this config to supply a {{meeting_topic}} variable and override the first message:

{
  "agent_id": "agent_YOUR_AGENT_ID",
  "dynamic_variables": {
    "meeting_topic": "Product planning"
  },
  "conversation_config_override": {
    "agent": {
      "first_message": "Hi, I'm ready to help with product planning."
    }
  }
}

Define the variable in your ElevenLabs prompt or messages, and enable the first-message override in the agent's security settings before using this example. See ElevenLabs' dynamic variables and overrides guides.

Recall.ai adds bot context to dynamic_variables and sends it with any conversation_config_override in the conversation_initiation_client_data message. If you omit overrides, the agent's saved configuration applies. conversation_config_override.conversation.text_only cannot be true, because the meeting integration requires audio.

Use Recall bot context in tools

Recall.ai automatically supplies these dynamic variables when starting the conversation:

VariableValue
recall_bot_idThe Recall bot's ID, matching the ID returned by Create Bot.
recall_api_base_urlThe bot's regional Recall API base URL, without an API version path, such as https://us-west-2.recall.ai.

Use {{recall_bot_id}} and {{recall_api_base_url}} in ElevenLabs fields that support dynamic variables, including your server tool configuration. Your tool server can use them to act on the meeting bot. It still needs your Recall API key to authenticate Recall API requests; the variables contain no Recall credentials.

These variables are supplied even when you omit dynamic_variables. Their names are reserved: Recall.ai replaces any values you supply for them and preserves your other variables. You do not need to add them to the bot request.

Append the endpoint's versioned path to this base URL when making Recall API requests.

Add a camera or avatar

ElevenLabs is supported under output_media.audio. You can combine it with an independent camera or screenshare source. When audio is configured, the meeting audio mix goes only to the audio agent; camera and screenshare sources do not receive it.

To use an automatic avatar, combine output_media.audio with automatic_video_output, leaving output_media.camera and output_media.screenshare absent.

FAQ

Where should I implement tool calls?

Use ElevenLabs server tools, such as webhook tools that call your application. Recall.ai does not execute client-side tools. It emits a customer-visible warning and returns an error result when the tool expects a response, while the conversation continues.

To add chat messages, screensharing, and a leave action, see Add meeting tools.


Did this page help you?